AI-driven robot hand spent hundred years teaching itself to rotate cube

Yet another example of the value of simulated training, researchers at OpenAI trained a robotic hand using hundreds of years of  object manipulation inside a computer simulation. The resulting hand movements are significantly more dexterous than anything developed so far.
https://rob.al/2vAOmBs
A reinforcement-learning algorithm allows Dactyl to learn physical tasks by practicing them in a virtual-reality environment.